Author’s Accepted Manuscript

Animal-assisted intervention for individuals with

cognitive impairment: a meta-analysis of
randomized controlled trials and quasi-randomized
controlled trials

Mingyue Hu, Ping Zhang, Minmin Leng, Chen Li,

Li Chen
www.elsevier.com/locate/psychres

PII: S0165-1781(17)30916-2
DOI: https://doi.org/10.1016/j.psychres.2017.12.016
Reference: PSY11052
To appear in: Psychiatry Research
Received date: 24 May 2017
Revised date: 10 November 2017
Accepted date: 8 December 2017
Cite this article as: Mingyue Hu, Ping Zhang, Minmin Leng, Chen Li and Li
Chen, Animal-assisted intervention for individuals with cognitive impairment: a
meta-analysis of randomized controlled trials and quasi-randomized controlled
trials, Psychiatry Research, https://doi.org/10.1016/j.psychres.2017.12.016
This is a PDF file of an unedited manuscript that has been accepted for
publication. As a service to our customers we are providing this early version of
the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting galley proof before it is published in its final citable form.
Please note that during the production process errors may be discovered which
could affect the content, and all legal disclaimers that apply to the journal pertain.
Animal-assisted intervention for individuals with cognitive impairment: a
meta-analysis of randomized controlled trials and quasi-randomized controlled trials

Mingyue Hua,1·Ping Zhanga,1·Minmin Lenga·Chen Lia·Li Chena,b*

a
College of Nursing, Jilin University, Changchun, China
b
Department of Pharmacology, college of Basic Medical sciences, Jilin University, Changchun,
China
Corresponding Author: Li Chen1,2, Department of Nursing, JiLin University, 965 XinJiang Road,
Changchun 130021, China. Tel +86 43185619366, Fax +86 43185619366
Email: chen__care@126.com.
1
These authors contributed equally to this work.

Abstract
The aim of this meta-analysis was to systematically examine the efficacy of
animal-assisted intervention (AAI) for cognitive impairment patients. PubMed,
Embase, the Cochrane library, PsycINFO, and Web of Science databases were
searched up to June 2017 to collect studies related to AAI conducted in patients with
cognitive impairment. Five randomized controlled trials (RCTs) and five
quasi-randomized controlled trials (quasi-RCTs) involving 413 participants were
included. Compared with control groups, AAI groups exhibited significantly fewer
behavioral and psychological symptoms of dementia (BPSD), especially depression
and agitation. In both the short and long term, AAI had beneficial effects on BPSD in
cognitive impairment patients. However, no significant improvements were found in
daily living activities, quality of life or cognitive score. The present meta-analysis
showed that AAI can be effective in reducing BPSD in patients with cognitive
impairment.
Keywords: animal-assisted intervention, AAI, cognitive impairment, dementia

1. Introduction
With the accelerating shift toward an aging global population, age-related
diseases have become a major focus of public health issues. According to World
Alzheimer Reports 2015 and 2016 (Prince et al., 2016; Prince et al., 2015), dementia
affects almost 50 million people globally, with a new case of dementia occurring
somewhere in the world every 3 seconds; this number is projected to increase to more
than 131 million by 2050. Furthermore, due to the insidious onset of cognitive
function decline, only 40-50% of those living with dementia in high-income countries
have received a diagnosis, a rate that is unlikely to exceed 5-10% in low- and
middle-income countries (Prince et al., 2015). There are over 100 forms of dementia,
of which the best known is Alzheimer’s disease (AD), accounting for 50-60% of all
cases (Alzheimer’s Disease International, 2017). Although there is currently no cure
for most types of dementia, treatment and support are available.
Behavioral and psychological symptoms of dementia (BPSD) are defined as
signs and symptoms of disturbed behavior, mood, thought or perception that
frequently occur in all stages of dementia but generally become more prevalent as
severity increases (Finkel et al., 1997; Rosenberg et al., 2015). The most common
BPSD in AD are agitation and depression; other BPSD include aggression, sleep
problems, and a variety of socially inappropriate behaviors (Kales et al., 2015;
Lyketsos et al., 2011; Rosenberg et al., 2015). Overall, BPSD are associated with
more rapid progression from mild AD to severe dementia and/or death, and they lead
to a heavy caregiver burden, distress, decreased quality of life and reduced well-being
(Givens et al., 2015; Peters et al., 2015; Wetzels et al., 2010).
Due to the adverse side effects of pharmacological therapies for treating BPSD
(Bierman et al., 2007; Russ, 2014), efforts are focusing on less risky
non-pharmacological interventions. A recent systematic review (Abraha et al., 2017)
showed that various non-pharmacological interventions, such as music therapy and
behavioral management techniques, are effective at reducing BPSD. Among the many
non-pharmacological interventions that exist, AAI (animal-assisted intervention) is a
prescribed, goal-directed approach aimed at ‘training’ the individual in one or more
specific abilities, and the results are promising (Abraha et al., 2017; Edwards and
Beck, 2002; Morrison, 2007). Indeed, studies since 1944 have shown that pets and
humans might have beneficial relationships for each (Bossard, 1944). In March 2013,
the International Association of Human-Animal Interaction Organizations (IAHAIO)
established AAI as a goal-oriented and structured intervention that intentionally
includes or incorporates animals in health, education, and human services for the
purpose of therapeutic gains in humans. AAIs incorporate human–animal teams in
formal human service such as animal-assisted therapy (AAT), animal-assisted
education (AAE), or, under certain conditions, animal-assisted activity (AAA)
(Jegatheesan et al., 2015). In addition to AAI’s underlying positive effects on
cognition, there is also research demonstrating that AAI has positive effects in many
other domains, such as the cardiovascular system, mood management, pain perception,
social contact, autism spectrum disorder, and general health (Barker et al., 2003;
Headey, 2008; McCune et al., 2017; Motooka et al., 2006). The underlying
mechanism can be explained by increased activation of the oxytocin system when
people are in contact with animals (Handlin et al., 2012; Odendaal and Meintjes,
2003). Moreover, the theory of anthropomorphism interprets the undesirable tendency
of humans to treat companion animals in a manner similar to humans as an efficient
way to provide social support and communication (Urquiza-Haas and Kotrschal,
2015). Other theories such as biophilia, attachment, verbal-symbolic system,
motivation, and social support may also be invoked to interpret the effects (Barker et
al., 2003; Beetz, 2017; Heinrichs et al., 2003).
A previous study showed that agitated behaviors significantly decreased while
social interaction significantly increased in dementia patients after AAI (Richeson,
2003). Conversely, another study indicated that although dementia patients had a
good impression of AAI, there was no significant difference in BPSD (Motomura et
al., 2004). In a recent review, the author included 18 articles to assess the effects of
AAI on elderly patients with dementia or psychiatric disorders, concluding that AAI
can reduce agitation and improve social interaction. Nonetheless, due to insufficient
data, there were no conclusions regarding effects of AAI on cognitive functions.
Although that review described encouraging results for AAI, a meta-analysis was not
involved, and the types of included studies were diverse. Moreover, new studies have
since emerged (Bernabei et al., 2013). Overall, some results of AAI have shown
promise, but no definite conclusion has been reached. Hence, the value of AAI in
those with cognitive impairment remains to be further investigated. The aim of the
present review was to conduct a meta-analysis of all studies completed thus far. The
study design encompassed randomized controlled trials (RCTs) and quasi-randomized
controlled trials (quasi-RCTs) that aimed to quantify the effect of AAI across patients
with cognitive impairment.
2. Method
2.1 Literature research
Although we were unable to register the protocol, we still followed the Preferred
Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines
(Moher et al., 2009). A comprehensive literature search in the electronic PubMed,
Embase, the Cochrane Library, PsycINFO, and Web of Science databases was
performed by two reviewers (HMY, LMM). To include all possible publications, we
searched all these databases up to June 2017. The language was restricted to English,
but the publication data were not restricted. The key search terms utilized a
combination of MeSH and entry terms (using the PubMed query as an example):
"Cognitive Dysfunction"[Mesh] OR "Cognitive Impairments" OR "Cognitive
Impairments" OR "Mild Neurocognitive Disorder" OR "Neurocognitive Disorder"
OR "Mild Cognitive Decline" OR "Cognitive Disorders" OR "Mental Disorder"[Mesh]
OR "Mental Deterioration" OR "Behavior Disorders" OR "Dementia"[Mesh] OR
"Amentia" OR "Alzheimer’s disease" OR "MCI" OR "Senile Paranoid Dementia" OR
"Familial Dementia", "Animal Assisted Therapy"[Mesh] OR "Pet Therapy" OR
"Assisted Therapy, Animal" OR " Animal Facilitated Therapy" OR "Therapies, Pet
Facilitated" OR "Animal Assisted Intervention" OR "animal-assisted activity",
"Randomized Controlled Trial"[Mesh], "Controlled Clinical Trial" OR "Randomized"
OR "Randomized Controlled Study" OR "Randomly" OR "Trial". Similar strategies
were used to search other databases.
2.2 Inclusion and exclusion criteria
Two reviewers (HMY, LMM) independently reviewed the list of potential
articles produced by the search strategy. Criteria for inclusion and exclusion were
determined a priori. The PICOS approach was used for eligibility criteria. The
inclusion criteria were as follows. (1) Study participants were diagnosed with
cognitive dysfunction, such as mild cognitive impairment, dementia, or Alzheimer’s
disease. (2) The intervention method was AAI. (3) The study was an RCT or a
quasi-RCT. We included quasi-RCTs because we had anticipated that there would be
a dearth of evidence from RCTs. (4) The study results reported details on the effects
of the intervention, or detailed data could be obtained by contacting the corresponding
authors. The exclusion criteria were as follows. (1) The study did not meet all the
PICOS standards. (2) There was no control group. (3) Cognitive dysfunction was
caused by other psychiatric or neurological disease, such as depression, stroke, and
accompanying medical disorders. (4) The publication only reported the protocol.
Missing data regarding outcomes of interest were obtained by contacting the
corresponding authors.
2.3 Study selection
After duplicate studies were eliminated (EndNote X7), titles were first reviewed
for obvious exclusions. After reading the abstract, the full text was reviewed if it was
unclear whether the article should be excluded. Two reviewers (HMY, LMM)
selected articles and extracted data independently and then compared their lists of
articles to ensure that the same studies had been included or excluded. The two
reviewers resolved any discrepancy by discussion. If discrepancies could not be
resolved, a third person (ZP) joined the discussion on how to classify the articles.
2.4 Data extraction
Two reviewers (HMY, LMM) independently extracted data from the included
studies and completed an a priori data table. The table of extracted data was reviewed
by a third person (ZP) to ensure that there were no errors. A standardized extraction
form was created to record the following data from the studies: author and year, types
of analysis, population, intervention type, intervention methods, allocation, mean age,
intervention format, follow-up times, frequency of intervention, drop-out, assessment
for level of cognitive impairment and outcome measures. The primary outcome was
BPSD; secondary outcomes were activities of daily living, cognitive state and quality
of life.
2.5 Quality assessment
The quality of the RCTs in six areas of bias known to affect clinical outcomes
(selection bias, performance bias, detection bias, attrition bias, reporting bias and
other bias) was assessed using the “Risk of Bias Tool” recommended by Cochrane
Handbook for Systematic Reviews of Interventions (Higgins and Green, 2011). The
methodological index for non-randomized studies (MINORS) was applied to further
assess the quality of each non-randomized study. The maximum global score for
comparative studies is 24 (Slim et al., 2003). All included studies were independently
assessed, and scores were determined via consensus.
2.6 Statistical analysis
The SMD (standardized mean difference) was calculated if there was variation in
the scale used by the trials. We pooled the results of trials using a fixed-effects model
when the heterogeneity test showed p > 0.05 and I2 < 50%. Otherwise, a
random-effects model and subgroup analysis were used to analyze the sources of
heterogeneity; the weight given to each study was based on the inverse of the variance.
All reported probabilities (p values) were two-sided, with p < 0.05 considered to be
statistically significant.
3. Results
3.1. Literature search and study characteristics
A total of 975 articles were identified from the four databases; 102 articles
remained after duplicate documents were removed and titles or abstracts were
reviewed. After reviewing the full text, 10 articles were included in the meta-analysis.
The selection process is shown in Fig. 1. Among the 10 articles, five each are RCTs
(Bono et al., 2015; Friedmann et al., 2015; Menna et al., 2016; Olsen et al., 2016a;
Olsen et al., 2016b) and quasi-RCTs (Churchill et al., 1999; Kanamori et al., 2001;
Majic et al., 2013; Moretti et al., 2011; Nordgren and Engström, 2014b). The detailed
characteristics of the included studies are presented in Table 1. Most of the
participants had a diagnosis of dementia, and dogs were used in most of the included
studies. The control groups in nine studies were exposed to the same conditions as the
corresponding intervention groups, except for the implementation of AAI in the latter.

However, in one study, the control group underwent reminiscence therapy, whereas
the intervention group received AAI. Figures on secondary outcomes are not
displayed.
Fig. 1. PRISMA flowchart of search and selection guidelines.

3.2. Study quality

We assessed the validity of individual trials using the Risk of Bias instrument.
This instrument is used to evaluate six key domains: selection bias (sequence
generation and allocation concealment); performance bias (blinding of participants,
personnel); detection bias (outcome assessors); attrition bias (incomplete outcome
data); reporting bias (selective outcome reporting); and other potential bias. We
excluded one of the studies from the meta-analysis because the rate of drop-out was
too high (Nordgren and Engström, 2014b). Detailed information is presented in Table
2.
3.3 Effectiveness of AAI in cognitive impairment patients
3.3.1 BPSD
All seven included studies (Churchill et al., 1999; Friedmann et al., 2015;
Kanamori et al., 2001; Majic et al., 2013; Menna et al., 2016; Moretti et al., 2011;
Olsen et al., 2016b) provided detailed data of BPSD, and the results showed that AAI
had a significantly beneficial effect on psychological symptoms in patients with
cognitive impairment in both random-effects [SMD = -1.23, 95%CI (-2.12 to -0.34), p
= 0.007] and fixed-effects [SMD = -0.84, 95%CI (-1.12 to -0.55), p＜0.00001]
models. There was substantial evidence of high heterogeneity (p＜0.05, I2 = 88%).
Only the result of the random-effects model is shown in Fig. 2.

Fig. 2. Effect of AAI on BPSD.

3.3.1.1 Subgroup analysis of BPSD

The results from four (Churchill et al., 1999; Friedmann et al., 2015; Majic et al.,
2013; Olsen et al., 2016b) studies showed that AAI had a significantly beneficial
effect on depression [SMD = -1.39, 95%CI (-2.17 to -0.08), p = 0.04]. In addition,
four (Friedmann et al., 2015; Majic et al., 2013; Moretti et al., 2011; Olsen et al.,
2016b) included studies showed a beneficial effect on agitation [SMD = -1.08, 95%CI
(-2.03 to -0.13), p = 0.03].
Fig. 3. Effect of AAI on agitation and depression.
The results from the seven (Churchill et al., 1999; Friedmann et al., 2015;
Kanamori et al., 2001; Majic et al., 2013; Menna et al., 2016; Moretti et al., 2011;
Olsen et al., 2016b) included studies showed no significant difference between
experimental and control groups at baseline (p = 0.18). Both short-term (Churchill et
al., 1999; Friedmann et al., 2015; Kanamori et al., 2001; Majic et al., 2013; Moretti et
al., 2011; Olsen et al., 2016b) and long-term (Friedmann et al., 2015; Menna et al.,
2016) AAI had a positive effect on psychological symptoms in cognitive impairment
patients [SMD = -0.96, 95%CI (-1.82 to -0.09), p = 0.03], [SMD = -3.15, 95%CI
(-3.95 to -2.35), p＜0.00001].

Fig. 4. Short-term and long-term effects of AAT on psychological symptoms.

3.3.2 Daily living activities

Three (Bono et al., 2015; Friedmann et al., 2015; Kanamori et al., 2001)
included studies provided detailed data on activities of daily living. The results
showed that AAI had no significant effect on daily living activities in fixed-effects
[SMD = 0.26, 95%CI (-0.18 to -0.71), p = 0.24] and random-effects [SMD = 0.24,
95%CI (-0.31to 0.79), p = 0.39] models. Heterogeneity was low (p = 0.22, I2 = 34%).
3.3.3 Cognitive state
Three (Kanamori et al., 2001; Menna et al., 2016; Moretti et al., 2011) included
studies provided detailed cognitive state data. Despite no significant effect on
cognitive state in a random-effects model [SMD = 1.56, 95%CI (-0.61 to 3.74), p =
0.16], a fixed-effects model [SMD = 0.68, 95%CI (0.09 to 1.26), p = 0.02] showed a
significant difference for AAI. Heterogeneity was high (p＜0.05, I2 = 92%).
3.3.4 Quality of life
Two (Olsen et al., 2016a; Olsen et al., 2016b) included studies provided quality
of life data. However, the results showed that AAI had no significant effect on quality
of life in a fixed-effects [SMD = 0.06, 95%CI (-0.34~0.47), p = 0.76] or a
random-effects [SMD = 0.06, 95%CI (-0.49 to 0.62), p = 0.82] model. Heterogeneity
was low (p = 0.17, I2 = 47%).
4. Discussion
Ten studies of AAI in cognitive impairment patients were included in our
meta-analysis. The results showed a statistically significant decrease in BPSD in
cognitive impairment patients treated with AAI, and subgroup analysis showed that
the effect included both depression and agitation. Short-term and long-term AAI both
had beneficial effects on BPSD, but we could not conclude that there were significant
improvements in activities of daily living, quality of life or cognitive state. There was
substantial evidence of high heterogeneity in BPSD and cognitive state (p＜0.05, I2＞
50%), which might be due to the severity of cognitive impairment or the different
assessment scales, as well as to other causes (Bernabei et al., 2013; Thodberg et al.,
2016). However, because of the mixture of patients with different levels of cognitive
impairment, we could not conduct subgroup analysis according to dementia severity.
Both random-effects and fixed-effects models showed that AAI had a significant
beneficial effect on psychological symptoms (both p＜0.05), but no significant effect
on daily living activities (both p＞0.05) or quality of life (both p＞0.05) was found.
Regarding cognitive state, random-effects and fixed-effects models yielded different
results, with the latter showing a significant difference between AAI and control
groups, which might be due to the different levels of disease severity or the potential
positive effect of AAI on cognitive state; we did not consider the former because of
the high heterogeneity. In summary, both random-effects and fixed-effects models
agreed on most outcomes, increasing the reliability of the results. The types of control
conditions might also be the original source of heterogeneity. For example, the control
groups in nine studies were exposed to the same conditions as the corresponding
intervention groups, with the exception of AAI. However, in one study (Friedmann et
al., 2015), the control group received reminiscence therapy while the intervention
group received AAI; when we excluded this study, the results still showed that AAI
had a significant beneficial effect on BPSD [SMD = -0.89, 95%CI (-1.66 to -0.12), p
= 0.02] and agitation [SMD = -0.53, 95%CI (-0.88 to -0.17), p=0.004], as well as
depression [SMD = -0.51, 95%CI (-0.87 to -0.17), p = 0.006]. In conclusion, these
data indicate that the results are highly stable.
In addition to many studies confirming the beneficial effect of AAI on BPSD
(Mossello et al., 2011; Nordgren and Engström, 2012), one early study suggested that
dog-assisted intervention can produce a substantial decrease in noise levels in an
experimental group when compared with a control group, which indicated a reduction
in verbal agitation in the experimental group (Walsh et al., 1995). Thus, along with
the frequently used outcome indicator of BPSD, programs that already utilize AAT
for patients with AD can also consider general ward noise as an outcome for
measuring the efficiency of AAI. In line with this finding, a study reported that total
agitated behavior was successfully modified by AAI (Sellers, 2006), and a
meta-analysis showed AAI to be associated with fewer depressive symptoms (Souter
and Miller, 2007). Some theories/mechanisms can help, especially those of us who are
managers and handlers of AAI programs, to achieve a better understanding of why
AAI can benefit individuals with cognitive deficit. Self-efficacy can explain the
phenomenon of motivation for behaviors, as AAI increases motivation for physical
activity in dementia patients, resulting in a decrease in BPSD (Bandura, 1977). The
results can also be interpreted in terms of anthropomorphism, biophilia, attachment,
social support, or other explanations (Barker et al., 2003; Beetz, 2017; Urquiza-Haas
and Kotrschal, 2015). In our study, subgroup analysis showed that both short-term
and long-term interventions had a beneficial effect on BPSD, which might suggest
that short-term intervention can be effective. Consequently, we propose that an
organization that is planning to initiate an AAI program should consider three months
as a time point for assessing whether AAI is an effective method in the elderly while
employing relevant measurements. Nonetheless, although an effect can be observed in
a one-month or another duration, we merely propose three months as a time point,
rather than baseline, for evaluating the validity of AAI. Two of the studies (Motomura
et al., 2004; Zisselman et al., 1996) found no significant improvement in BPSD,
activities of daily living, or cognitive state in relation to AAI; however, the visit times
were both less five days, which might be too short to reach a significant result. Our
study defined the maximum length of a short-term intervention as three months.
Similarly, a recent study (Thodberg et al., 2016) found a decrease in depression
during the experimental period but not the entire process, which might be attributable
to the low intensity of the 10-minute dog visits. Based on these results, when the
worker of an organization is going to carry out AAI, it is suggested that the length of
intervention time be not less than five days and that the time per session of AAI
should greater than 10 minutes. Clearly, we can be able to explore the efficiency of
AAI effectively on the premise that the duration, length and time per session are
adequate.
Nonetheless, we did not conclude that there were significant improvements in
daily living activities, cognitive state or quality of life. Previous research involving a
geriatric population had shown an increased salivary cortisol level after interaction
with a dog, which suggested active involvement in daily living activities (Berry et al.,
2012). In contrast, many studies have reported lower levels of cortisol during contact
with animals (Barker et al., 2010; Handlin et al., 2012; Heinrichs et al., 2003), and
cortisol might be considered as a ‘stress response’ that was suppressed by social
support through AAI. Regardless, cortisol levels are easily affected by various types
of stimuli, such as movement and activity, which might be influential factors (Handlin
et al., 2012). Thus, further research should be performed to substantiate the
relationship between cortisol and activities involving dogs. Indeed, because the results
for associations between cortisol and AAI are mixed, it is our responsibility to
determine whether AAI has a positive impact on daily living activities, cognitive state
and quality of life and then further explore the role of cortisol in AAI. Our results are
not surprising because dementia is a progressive brain disease, and it would be
unrealistic to expect AAI to reverse the cognitive state to a normal state. Additionally,
there is evidence showing that patients with cognitive impairment had improved
quality of life after AAI (Nordgren and Engström, 2014a); however, due to the
insufficient evidence in these categories in our selected studies, larger studies are
needed to explore other parameters in addition to BPSD. We also suggest that
planners of AAI programs who are formulating or will formulate AAI protocols
should not only focus on outcomes of BPSD but also on daily living activities,
cognitive state, noise levels or quality of life, as these facets can serve as indicators
reflecting the physical and psychological conditions of the elderly.
Finally, our results suggest that AAI may be a suitable nonpharmacological
treatment for BPSD in cognitive impairment patients and should be considered as a
useful tool in clinical practice. There are several issues that need to be noted (safety,
applicability, time, format). First and foremost, the safety of elderly patients is of
paramount importance; therefore, the intervention should be applied carefully and
implemented by professional handlers. Second, applicability might vary from person
to person, as not all people are fond of animals or interested in AAI. Third, it would
be best to establish a formal guideline for the timing of AAI, including the duration of
the entire process and of each session as well as the frequency. In most of the studies
included in our analysis, the duration of the entire intervention, the time per session,
and the frequency ranged from six weeks to six months (three and six months were
common), 30 minutes to 90 minutes (30 and 45 minutes were common), and once
every two weeks to twice a week (mostly once a week), respectively. Intervention
formats involving groups and individuals accounted for 60% and 40%, respectively,
of the studies in our analysis, and we recommend that the internal and external
environments should be taken into consideration when choosing intervention formats.
In addition to AAIs, recent studies based on certain nonliving substitutes yielded
positive results; for instance, doll therapy was effective and economical in promoting
and maintaining caregiving and exploration abilities in dementia patients (Pezzati et
al., 2014; Ng et al., 2017). Other substitutes such as robots may also be promising
options for the future, as dementia patients exposed to a robotic toy that mimicked
either a dog or a baby showed decreased symptoms of BPSD (Jøranson et al., 2015;
Tamura et al., 2004). Future research should include more robust randomized
controlled trials to verify the effectiveness of the intervention.

5. Limitation
Our meta-analysis is limited in that we searched only the English literature. Thus,
some studies may have been missed. Due to the dearth of evidence from RCTs, we
also included quasi-RCTs. The small sample size included is another main limitation,
but we believe that with the development of non-pharmacological interventions, an
increasing number of researchers will explore the effectiveness of AAI in cognitive
impairment patients.
Funding sources
This research did not receive any specific grants from funding agencies in the
public, commercial, or not-for-profit sectors.
Conflict of interest
The authors declare no conflicts of interest.
Acknowledgment
 We thank all the authors of the included papers.

References

Abraha, I., Rimland, J.M., Trotta, F.M., Dell'Aquila, G., Cruz-Jentoft, A., Petrovic, M., et al., 2017. Systematic review of systematic

reviews of non-pharmacological interventions to treat behavioural disturbances in older patients with dementia. The

SENATOR-OnTop series. BMJ Open 7 (3), e012759.

Alzheimers’s Disease International (ADI), 2017. About dementia. https://www.alz.co.uk/about-dementia (accessed 17.06.03).

Bandura, A., 1977. Self-efficacy: toward a unifying theory of behavioral change. Psychol. Rev. 84 (2), 191.

Barker, S.B., Knisely, J.S., McCain, N.L., Schubert, C.M., Pandurangi, A.K., 2010. Exploratory Study of StressBuffering Response

Patterns from Interaction with a Therapy Dog. Anthrozoös 23 (1), 79-91.

Barker, S.B., Pandurangi, A.K., Best, A.M., 2003. Effects of animal-assisted therapy on patients' anxiety, fear, and depression before ECT.

J. ECT 19 (1), 38-44.

Beetz, A.M., 2017. Theories and possible processes of action in animal assisted interventions. Appl. Dev. Sci. 21 (2), 139-149.

Bernabei, V., De Ronchi, D., La Ferla, T., Moretti, F., Tonelli, L., Ferrari, B., et al., 2013. Animal-assisted interventions for elderly

patients affected by dementia or psychiatric disorders: A review. J. Psychiatr. Res. 47 (6), 762-773.

Berry, A., Borgi, M., Terranova, L., Chiarotti, F., Alleva, E., Cirulli, F., 2012. Developing effective animal-assisted intervention

programs involving visiting dogs for institutionalized geriatric patients: a pilot study. Psychogeriatrics 12 (3), 143-150.

Bierman, E.J., Comijs, H.C., Gundy, C.M., Sonnenberg, C., Jonker, C., Beekman, A.T., 2007. The effect of chronic benzodiazepine use

on cognitive functioning in older persons: good, bad or indifferent? Int. J. Geriatr. Psych. 22 (12), 1194-1200.

Bono, A.V., BenVenuti, C., Buzzi, M., CiAtti, R., ChiARelli, V., ChiAMBRetto, P., et al., 2015. Effects of animal assisted therapy (AAT)

carried out with dogs on the evolution of mild cognitive impairment. G. Gerontol. 63, 32-36.

Bossard, J.H., 1944. The mental hygiene of owning a dog. Mental hygiene(28), 408-413.

Churchill, M., Safaoui, J., McCabe, B.W., Baun, M.M., 1999. Using a therapy dog to alleviate the agitation and desocialization of people

with Alzheimer's disease. J. Psychosoc. Nurs. Men. 37 (4), 16-22.

Edwards, N.E., Beck, A.M., 2002. Animal-assisted therapy and nutrition in Alzheimer’s disease. Western J. Nurs. Res. 24 (6), 697-712.

Finkel, S.I., e Silva, J.C., Cohen, G., Miller, S., Sartorius, N., 1997. Behavioral and psychological signs and symptoms of dementia: a

consensus statement on current knowledge and implications for research and treatment. Int. Psychogeriatr. 8 (S3), 497-500.

Friedmann, E., Galik, E., Thomas, S.A., Hall, P.S., Chung, S.Y., McCune, S., 2015. Evaluation of a pet-assisted living intervention for

improving functional status in assisted living residents with mild to moderate cognitive impairment: a pilot study. Am. J. Alzheimers

Dis. 30 (3), 276-289.

Givens, J.L., Jones, R.N., Mazor, K.M., Prigerson, H.G., Mitchell, S.L., 2015. Development and psychometric properties of the family

distress in advanced dementia scale. J. Am. Med. Dir. Assoc. 16 (9), 775-780.

Handlin, L., Nilsson, A., Ejdebäck, M., Hydbring-Sandberg, E., Uvnäs-Moberg, K., 2012. Associations between the Psychological

Characteristics of the Human–Dog Relationship and Oxytocin and Cortisol Levels. Anthrozoös 25 (2), 215-228.

Headey, B., Na, F., & Zheng, R., 2008. Pet dogs benefit owners’ health: A “natural experiment” in china. Soc. Indic. Res. 84, 481–493.

Heinrichs, M., Baumgartner, T., Kirschbaum, C., Ehlert, U., 2003. Social support and oxytocin interact to suppress cortisol and subjective

responses to psychosocial stress. Biol. Psychiat. 54 (12), 1389-1398.

Higgins, J.P.T., Green, S., 2011. Cochrane Handbook for systematic Reviews of Interventions Version 5.1.0. The Cochrane

Collaboration.

Jøranson, N., Pedersen, I., Rokstad, A.M.M., Ihlebæk, C., 2015. Effects on symptoms of agitation and depression in persons with

dementia participating in robot-assisted activity: a cluster-randomized controlled trial. J. Am Med. Dir. Assoc. 16 (10), 867-873.

Jegatheesan, B., Beetz, A., Ormerod, E., Johnson, R., Fine, A.H., Yamazaki, K., et al., 2015. The IAHAIO Definitions for

Animal-Assisted Intervention and Guidelines for Wellness of Animals Involved. Handbook on animal-assisted therapy, 415-418.
Kales, H.C., Gitlin, L.N., Lyketsos, C.G., 2015. Assessment and management of behavioral and psychological symptoms of dementia.

BMJ 350, h369.

Kanamori, M., Suzuki, M., Yamamoto, K., Kanda, M., Matsui, Y., Kojima, E., et al., 2001. A day care program and evaluation of

animal-assisted therapy (AAT) for the elderly with senile dementia. Am. J. Alzhemers Dis. 16 (4), 234-239.

Lyketsos, C.G., Carrillo, M.C., Ryan, J.M., Khachaturian, A.S., Trzepacz, P., Amatniek, J., et al., 2011. Neuropsychiatric symptoms in

Alzheimer’s disease. Alzheimers Demen. 7 (5), 532 -539.

Majic, T., Gutzmann, H., Heinz, A., Lang, U.E., Rapp, M.A., 2013. Animal-assisted therapy and agitation and depression in nursing

home residents with dementia: a matched case-control trial. Am. J. Geriatr. Psychiat. 21 (11), 1052-1059.

McCune, S., Esposito, L., Griffin, J.A., 2017. Introduction to a thematic series on animal assisted interventions in special populations.

Appl. Dev. Sci. 21 (2), 136-138.

Menna, L.F., Santaniello, A., Gerardi, F., Di Maggio, A., Milan, G., 2016. Evaluation of the efficacy of animal-assisted therapy based on

the reality orientation therapy protocol in Alzheimer's disease patients: a pilot study. Psychogeriatrics 16 (4), 240-246.

Moher, D., Liberati, A., Tetzlaff, J., Altman, D.G., Group., P., 2009. Preferred reporting items for systematic reviews and meta-analyses:

the PRISMA statement. Plos med. 6 (7), e1000097.

Moretti, F., De Ronchi, D., Bernabei, V., Marchetti, L., Ferrari, B., Forlani, C., et al., 2011. Pet therapy in elderly patients with mental

illness. Psychogeriatrics 11 (2), 125-129.

Morrison, M.L., 2007. Health Benefits of Animal-Assisted Interventions. Complementary Health Practice Review 12 (1), 51-62.

Mossello, E., Ridolfi, A., Mello, A.M., Lorenzini, G., Mugnai, F., Piccini, C., et al., 2011. Animal-assisted activity and emotional status

of patients with Alzheimer's disease in day care. Int. Psychogeriatr. 23 (6), 899-905.

Motomura, N., Yagi, T., Ohyama, H., 2004. Animal assisted therapy for people with dementia. Psychogeriatrics 4 (2), 40-42.

Motooka, M., Kennedy, N.L., Koike, H., Yokoyama, T., 2006. Effect of dog-walking on autonomic nervous activity in senior citizens.

Med. J. Austrilia 184 (2), 60-63.

Ng, Q.X., Ho, C.Y., Koh, S.S., Tan, W.C., Chan, H.W., 2017. Doll therapy for dementia sufferers: A systematic review. Complement

Ther. Clin. Pract. 26, 42-46.

Nordgren, L., Engström, G., 2012. Effects of animal-assisted therapy on behavioral and/or psychological symptoms in dementia: A case

report. m. J. Alzheimers Dis. 27 (8), 625-632.

Nordgren, L., Engström, G., 2014a. Animal-assisted intervention in dementia: effects on quality of life. Clin. Nurs. Res. 23 (1), 7-19.

Nordgren, L., Engström, G., 2014b. Effects of dog-assisted intervention on behavioural and psychological symptoms of dementia: Lena

Nordgren and Gabriella Engström present the results from a study that used dogs as an alternative or a complement to

pharmacological treatments. Nursing older people 26 (3), 31-38.

Odendaal, J.S.J., Meintjes, R.A., 2003. Neurophysiological Correlates of Affiliative Behaviour between Humans and Dogs. Vet. J. 165

(3), 296-301.

Olsen, C., Pedersen, I., Bergland, A., Enders-Slegers, M.J., Ihlebaek, C., 2016a. Effect of animal-assisted activity on balance and quality

of life in home-dwelling persons with dementia. Geriatr. Nurs. 37 (4), 284-291.

Olsen, C., Pedersen, I., Bergland, A., Enders-Slegers, M.J., Patil, G., Ihlebaek, C., 2016b. Effect of animal-assisted interventions on

depression, agitation and quality of life in nursing home residents suffering from cognitive impairment or dementia: a cluster

randomized controlled trial. Int. J. Geriatr. Psych. 31 (12), 1312-1321.

Peters, M.E., Schwartz, S., Han, D., Rabins, P.V., Steinberg, M., Tschanz, J.T., et al., 2015. Neuropsychiatric symptoms as predictors of

progression to severe Alzheimer's dementia and death: the Cache County Dementia Progression Study. Am. J. Psychiat. 172 (5),

460-465.

Pezzati, R., Molteni, V., Bani, M., Settanta, C., Di Maggio, M.G., Villa, I., et al., 2014. Can Doll therapy preserve or promote attachment

in people with cognitive, behavioral, and emotional problems? A pilot study in institutionalized patients with dementia. Front

Psychol 5, 342.

Prince, M., Comas-Herrera, A., Knapp, M., Guerchet, M., Karagiannidou, M., 2016. World Alzheimer report 2016: improving healthcare
 for people living with dementia: coverage, quality and costs now and in the future. Alzheimer's Disease International.

Prince, M., Wimo, A., Guerchet, M., Ali, G.C., Wu, Y.T., Prina, M., 2015. World Alzheimer Report 2015: The global impact of dementia.

An analysis of prevalence, incidence, cost & trends; Alzheimer’s Disease International: London. Alzheimer's Disease International.

Richeson, N.E., 2003. Effects of animal-assisted therapy on agitated behaviors and social interactions of older adults with dementia. . Am.

J. Alzheimers Dis. 18 (6), 353-358.

Rosenberg, P.B., Nowrangi, M.A., Lyketsos, C.G., 2015. Neuropsychiatric symptoms in Alzheimer's disease: What might be associated

brain circuits? Mol. Aspects Med. 43-44, 25-37.

Russ, T.C., 2014. Cholinesterase inhibitors should not be prescribed for mild cognitive impairment. Evid-based. Med. 19 (3).

Sellers, D.M., 2006. The Evaluation of an Animal Assisted Therapy Intervention for Elders with Dementia inLong-Term Care. Activities,

Adaptation & Aging 30 (1), 61-77.

Slim, K., Nini, E., Forestier, D., Kwiatkowski, F., Panis, Y., Chipponi, J., 2003. Methodological index for non‐ randomized studies

(MINORS): development and validation of a new instrument. ANZ J. surg. 73 (9), 712-716.

Souter, M.A., Miller, M.D., 2007. Do Animal-Assisted Activities Effectively Treat Depression? A Meta-Analysis. Anthrozoös 20 (2),

167-180.

Tamura, T., Yonemitsu, S., Itoh, A., Oikawa, D., Kawakami, A., Higashi, Y., et al., 2004. Is an entertainment robot useful in the care of

elderly people with severe dementia? J. Gerontol. A-biol. 59 (1), M83-M85.

Thodberg, K., Sørensen, L.U., Christensen, J.W., Poulsen, P.H., Houbak, B., Damgaard, V., et al., 2016. Therapeutic effects of dog visits

in nursing homes for the elderly. Psychogeriatrics 16 (5), 289-297.

Urquiza-Haas, E.G., Kotrschal, K., 2015. The mind behind anthropomorphic thinking: attribution of mental states to other species. Anim.

Behav. 109, 167-176.

Walsh, P.G., Mertin, P.G., Verlander, D.F., Pollard, C.F., 1995. The effects of a ‘pets as therapy’dog on persons with dementia in a

psychiatric ward. Aust. Occup. Ther. J. 42 (4), 161-166.

Wetzels, R.B., Zuidema, S.U., de Jonghe, J.F., Verhey, F.R., Koopmans, R.T., 2010. Determinants of quality of life in nursing home

residents with dementia. Dement. Geriatr. Cogn. Disord. 29 (3), 189-197.

Zisselman, M.H., Rovner, B.W., Shmuely, Y., Ferrie, P., 1996. A pet therapy intervention with geriatric psychiatry inpatients. Am. J.

Occup. Ther. 50 (1), 47-51.

Table 1. Characteristics of included studies.
Assessm
Mean Frequenc
ent of
age y, total
Interventi the level
Author, Nature of (mean ± Interven-t session, Drop
Populatio on types Types of Intervention Alloca-t Follow-u of Outcome
year, Setting intervention SD) ion duration -out
n for EG animal methods ion ps cognitiv measures
types s Sex format of (%)
and CG e
(female intervent
impairm
%) ion
ent
Mild, Day-care EG: Dogs Physical Greeting the EG: 42 - Group 3 Twice a 1.25 MMSE Berg Balance
moderate centers dog-assist tasks: dog, petting CG: 38 - sessions: week, 12 % Scale/Quality
, and ed To increase the dog, baseline, weeks, of Life in
severe interventi physical brushing the 3 30 min Late-Stage
dementia on activities by dog, feeding months, Dementia.
Olsen et
CG: bending the dog a 6 months
al.
routine down, treat, and
2016
care reaching throwing a
RCT
out, lifting toy for the
their arms, dog to fetch.
and
throwing a
ball.
Nordgre Moderate Nursing EG: Dogs: 1 Individualiz Walking or EG: 20 82 Individua 3 Once or 42% MMSE, Cohen-Mansfi
n et al. and homes validation Boxer; ed playing with CG: 13 - l sessions: twice a DSM-IV eld Agitation
2014 severe therapy, 1 Golden protocols: the dog, baseline, week, 10 Inventory/
quasi-R dementia massage, Retriever; To promote petting, the 3 sessions, Multi-Dimensi
 CT singing, 2 communicat dog, feeding months, 45-60 onal Dementia
reminisce Flatcoate ion with it treats, 6 minutes Assessment
nce d dogs and talking to it, months. Scale.
activities Retriever dog brushing it,
+ s. handlers. reminiscing
dog-assist about
ed previous
interventi pets, or
on talking to
CG: the dog
validation handler.
therapy,
massage,
singing,
reminisce
nce
activities
Moderate Nursing EG: Dogs: Physical, Speaking to EG: 35 81.8 Individua 2 Once a 13.33 MMSE Mini-Mental
and homes dog-assist Border verbal tasks: the dog, CG: 40 72.31% l sessions: week, 10 % State
severe ed Collies. To promote stroking and baseline, weeks, Examination/
Majic et
dementia interventi verbal, petting the 10 weeks 45 Cohen-Mansfi
al. 2013
on active dog, minutes eld Agitation
quasi-R
CG: interaction throwing a Inventory/
CT
routine with dogs. ball for the Dementia
care dog to Mood
retrieve, Assessment
 free Scale
interaction
Dementi Nursing EG: Dogs Different Petting the EG: 28 83.5 Group 3 Twice a 12% MMSE Norwegian
a, homes dog-assist activities: dog, feeding CG: 30 62.7% sessions: week, 3 versions of the
cognitive ed To increase the dog a baseline, months, Cornell Scale
deficit interventi physical treat, finishing 30 for
on activities by throwing a the minutes Depression/Br
CG: bending toy for the intervent ief Agitation
routine down, dog to fetch ion after Rating
Olsen et
care reaching 12 Scale/Quality
al. 2016
out, lifting weeks, of Life in
RCT
their arms, follow-u Late-stage
and p 3 Dementia
throwing a months scale
ball. after the
intervent
ion had
ended
Mild and Assisted-li EG: Dogs: Various -feeding the EG: 22 80.7 Group 2 Twice a 0.7% MMSE Barthel
moderate ving dog-assist Cardigan activities: dog, CG: 18 72.5% sessions: week, 12 Index/amount
Friedma dementia residents ed Welsh To increase brushing the baseline, weeks, of physical
nn et al. interventi Corgi. the dog’s teeth, 3 months 60-90 activity/
2015 on activities of brushing the minutes Cornell Scale
RCT CG: daily living, dog’s fur, for Depression
reminisce range of and dressing in
nce motion, fine the dog in a Dementia/7-It
therapy motor skills, bandana; em Apathy
ability of -throwing a Evaluation
sequencing ball, Scale/
events, and grooming Cohen-Mansfi
social skills. the dog; eld Agitation
-adjusting Inventory
the collar,
hand-feedin
g a treat,
and petting
the dog;
-opening a
container of
treats and
then giving
the dog a
treat;
-talking to
the dog,
talking
about the
dog to
another
person,
giving the
dog
 commands,
and petting
the dog.
Dementi Nursing EG: Dogs: 3 Some Holding, EG: 10 84.7 Group 2 Once a 5% MMSE Global
a, home dog-assist Golden activities: stroking, CG: 11 90.5% sessions: week, 6 Deterioration
cognitive ed Retriever To promote walking, baseline, weeks, Scale/
deficit interventi s; verbal, talking to 6 weeks 90 Mini-Mental
on 1 active and playing minutes State
CG: Pincher. interaction with the Examination
routine with dogs. animals. /Mini-Mental
Moretti care; State
et al. allowed Examination
2011 to see the
quasi-R animals
CT coming
into the
nursing
home but
not
allowed
to interact
with them
Churchi Dementi Extend-car EG: pet Dogs: To provide EC: 16 83.8 Group 1 hour Twice a 12.5 BDBRS Agitation
ll et al. a e facilities therapy Labrador participants CC: 16 65.6% day, 1 % Behavior
1999 interventi Retriever with better day, 30 Mapping
quasi-R on s. communicat minutes Instrument/
 CT CC: ion with Daubenmire’s
routine dogs. data coding
care protocol

Alzheim Day-care EC: dog Dogs and - - EC: 7 81.4 Individua 2 Once 0% MMSE Mini-Mental
er’s center and cat cats. CC: 12 89.47% l sessions: every State
Kanamo disease, therapy baseline, two Examination/
ri et al. vascular interventi 3 months weeks, Nishimura’s
2001 dementia on 12 ADL/Behavio
quasi-R CG: weeks, ral Pathology
CT routine in
care Alzheimer’s
Disease
Mild and Alzheimer’ EG: dog Dog: Structured Reintroduci EG: 16 75.2 Group 2 Once a 0% MMSE Global
moderate s daycare therapy Labrador activities: ng the dog CG: 7 75% sessions: week, 6 Deterioration
Alzheim center interventi retriever. To stimulate to each baseline, months, Scale/
er’s on cognitive patient and 6 months 45 mins Mini-Mental
disease CG: functions helping the State
Menna
routine including patient Examination
et al.
care attention, develop
2016
language spatial-temp
RCT
skills, and oral
spatial-temp orientation,
oral structured
orientation. activity with
the dog.
 Minor Alzheimer EG: dog Dogs: 2 Specific & - EG: 16 80.2 Individua 2 Once 25% MMSE Barthel Index
and mild evaluation therapy Border standard identificatio CG: 16 66.67% l sessions: every
AD unit interventi Collie; 1 activities: n of the dog baseline, two
on Golden To stimulate breed 8 months weeks, 8
CG: Retriever cognitive looking at months,
routine s; 1 functions, illustrations, 60
care Weimara communicat recalling of minutes
ner. ion with personal
peers, and events with
manual animals,
Bono et
skills. learning
al. 2015
basic
RCT
commands,
rememberin
g some play
with the
animal.
-learning
the body
language
of the
animal.
Abbreviations: RCT, randomized controlled trials; quasi-RCT: quasi-experimental studies; EG, experimental group; CG: control group. DSM-IV: Diagnostic and Statistical
Manual of Mental Disorders, Fourth Edition; BDBRS: Burke Dementia Behavioral Rating Scale.

Table 2. Quality assessment with the Cochrane Collaboration’s Risk of Bias Tool and MINORS.
 RCT quasi-RCT
Intention-t Incomplete Selective Other
Studies Sequence Allocation Blinding of Assessor MINORS
o-treat outcome outcome potential
generation concealment participants blinding score
analysis data reporting bias
Olsen et al. (2016) L L L H L L L L
Nordgren et al. (2014) 17
Majic et al. (2013) 21
Olsen et al. (2016) L L L U L L L L
Friedmann et al. (2015) L L L L L L L L
Moretti et al. (2011) 19
Churchill et al. (1999) 18
Kanamori et al. (2001) 21
Menna et al. (2016) H H U U L L L L
Bono et al. (2015) L L L L L L L L
Abbreviations: RCT, randomized controlled trials; quasi-RCT: quasi-experimental studies; L, low risk; H, high risk; U, unclear risk.

Hights:

 Using meta-analysis to examine the efficacy of Animal-assisted intervention (AAI) for cognitive impairment patients.

 To explore the effect of animal-assisted intervention on cognitive impairment patients.